Measurement Analysis Of Atmospheric Boundary Layer Turbulence Signal

The research of turbulence signal were widely paid attention to many fields, such as mechanical, meteorology, fluid fields. Especially, for the exploration and analysis of wind speed and the temperature in the atmosphere.Various methods of processing signal, which can deep help people research and solve all kinds of turbulent problems, have indispensable functions in the development process of turbulence research. However, the traditional signal analysis methods solving turbulent problems have many limit and maybe make some bias in the research. Thus, signal analysis methods should be improved in researching the turbulent problems. Based on Hilbert-Huang Transform, the paper analyzed temperature and wind speed signals in atmosphere, the specific researches are as follow:Firstly, analyzing inner changing rule of temperature and wind speed signal about the Xu Jiahui and Fengxian in Shanghai, which is based on HHT method. The results showed that the time series of fairly strong intermittent are in fairly large scale modes about bout temperature and wind speed signals; And only rarely modes obeyed Gaussian Distribution, most of modes deviate from the Gaussian Distribution. The coherent structures gotten from the empirical mode decomposition showed that wind speed and temperature signal in the atmosphere are controlled by large scale coherent structures.Secondly, the analysis of the temperature and wind speed signal showed that the part of the Hilbert marginal spectrum and Fourier spectrum follow the Kolmogorov’s law in inertia area. And the Hilbert spectrum showed that the power of temperature and speed fluctuations are in the low frequency areas. Combined with coherent structures, energy were contained in the coherent structures of large scale. The correlation of temperature and wind speed in Fengxian and Xu Jiahui showed that there are obvious correlation in both large and small time scale of their temperature. But, almost last two order modes and trends presented obvious correlation in both south-north and east-west wind speed.Then, In nearly 50 years, in the study of monthly average temperature and high (low) temperature, it found that temperature time series mainly follow the change law of oscillation for a year period; Modes of the monthly average temperature have a little bit strong intermittent; A certain time scale of the monthly highest and lowest temperatures time series have much stronger intermittent; Through the study of the highest temperature and minimum temperature days found that the changing rules of the signals are stable, and the change of the main period are in small time scale; Annual highest temperature days has strong intermittent mode, but the modes of annual minimum temperature days have a little bit intermittent.Finally, the prediction studies about three groups of signals of monthly mean temperature, monthly maximum temperature and minimum temperature showed that the RBF neural network prediction results based on the EMD decomposition of are more accurate than the simple RBF neural network prediction results. And in the study of short-term signals, the RBF neural network prediction results is not very accurate.